Rethinking user feedback

A Github topic about rethinking how user feedback works generated a lot of response, so I've decided to move the conversation here and open it up.

Motivation

Once in awhile someone complains to me about feedback they received. They want it purged, they think it's wrong, they think there's some conspiracy to target them.

This got me to ask, what is the purpose of negative user feedback? This sort of users-reviewing-users system is something I think is unique to Danbooru. The site offers plenty of mediums for communicating with other users (dmails, forums, comments). What utility does user feedback bring that these other mediums don't give?

Is the goal to critique, warn, and inform users of mistakes they've made?

Is the goal to warn others about the user, to keep a history of their behavior?

Is it to ostracize and shame the user?

Is it deterrence, to motivate other users to behave appropriately?

Once a feedback is made it's there and the user is forced to plead to mods to argue their case. There's no formal review process, no requirements for evidence.

Proposal

I initially proposed to limit the length of feedback. After reading the discussion on Github though I've changed my mind. I think what's needed instead is an explicit policy change about how negative feedback is given and how it can be adjudicated.

Here's what I propose:

The user is the person who received the feedback. The critic is the person who gives the feedback.

• Before giving negative (or even neutral) feedback, you must show a sincere effort at communicating with the user through Dmails.
• If after two weeks you see no change, you can then submit a negative feedback. You should include permalinks to the dmails and any relevant evidence. This includes stuff like searches, uploads, comments, etc. Notably, you can only cite things you know of first-hand. You cannot rely on testimony of others. They are free to offer their own feedback.
• If the user feels the feedback was given wrongly, they are free to open a forum topic disputing it, tagging a mod. If the critic is a mod, you can tag an admin.
• Historically I've been involved in conversations with disputes, but I think 1-on-1 is not the appropriate medium for this sort of dialogue. It's easy for me to miss context or evidence this way, and it's annoying to act as a mediator between two people via messages. It should be open for anyone to chime in.
• The user must explain why the stated evidence (if any is provided) is wrong. They should link to contrary evidence that proves their innocence.
• Burden of proof is ultimately on the critic. If they can't provide adequate evidence, or counter the user's defense, then the feedback will be rolled back by default.
• Ultimately, a mod or admin must make a judgement call about whether the feedback stands.
• If the user loses their dispute, a neutral feedback will be recorded indicating as such. Any future adjudicating mod or admin can decide whether future disputes are worth their time.

This is a lot of process. But I also feel like people shouldn't be handing out negative feedback willy-nilly without at least making some attempt at communication beforehand. Keep in mind neutral feedback would not be subject to any of this.

This is much more reasonable a change of policy than what was previously proposed. Several of us already try to contact the one we're issuing the feedback to unless it's clear that the person we're negging is a vandal or sockpuppet. In fact it's even mentioned that such is the best course of action in the feedback wiki. Making contacting the user in question necessary for the issuing of the feedback is a good idea, because any possible drawback (I can't really think of any, and sending a DMail only takes a couple of minutes) would be superseded by the fact that disputes can be solved amicably in the cases where users simply did not know better. I stand by my proposal that a notice should be displayed in the feedback form to make sure everyone is aware of this rule.

Making feedback removal/appeal public also is a very good way to approach the problem imo, because it makes sure that there can't be unwarranted abuse of power without it being public and subject to dispute (nobody really checks the mod actions, but a public forum thread is another story entirely). However such thing should be made mandatory - there shouldn't be cases where mods can singlehandedly remove other users' feedbacks like they did in the past, unless it's absolutely evident that it's plain vandalism (like what happened when Killias deleted his account and made a bunch of nonsensical feedbacks two months ago). This has already been abused in the past, even to remove feedbacks issued to oneself, until such ability was patched out from the site code, so the danger is very real.

nonamethanks said:

This is much more reasonable a change of policy than what was previously proposed. Several of us already try to contact the one we're issuing the feedback to unless it's clear that the person we're negging is a vandal or sockpuppet. In fact it's even mentioned that such is the best course of action in the feedback wiki. Making contacting the user in question necessary for the issuing of the feedback is a good idea, because any possible drawback (I can't really think of any, and sending a DMail only takes a couple of minutes) would be superseded by the fact that disputes can be solved amicably in the cases where users simply did not know better. I stand by my proposal that a notice should be displayed in the feedback form to make sure everyone is aware of this rule.

Making feedback removal/appeal public also is a very good way to approach the problem imo, because it makes sure that there cannot be unwarranted abuse of poewr without it being public and subject to dispute (nobody really checks the mod actions, but a public forum thread is another story entirely). However such thing should be made mandatory - there shouldn't be cases where mods can singlehandedly remove other users' feedbacks like they did in the past, unless it's absolutely evident that it's plain vandalism (like what happened when Killias deleted his account and made a bunch of nonsensical feedbacks two months ago). This has already been abused in the past, even to remove feedbacks issued to oneself, until such ability was patched out from the site code, so the danger is very real.

I agree with pretty much everything said here by NNT. This is a *much* better approach.

So negative feedbacks will have to be used the way they should be used anyway? I’m all for more hard facts that can’t be explained away as slander or conspiracies.

I see one big problem, though, and that is tampering with evidence. nonamethanks already touched on that above. If giving evidence is important, it needs to be reliable. Currently, some content can be expunged (double deleted) without a trace. For example, comments can be expunged and expunged comments are indistinguishable from comments that never existed in the first place. A dead comment link certainly doesn’t have much impact as evidence. Expunging content is desirable and warranted in some cases, such as PII, spam and malicious content like malware. However, it should leave some kind of trace. The content may be lost, which is the point of expunging something, but we’ll know that something has existed there and who removed it.

I think two simple changes should be implemented to allow resilient negative feedbacks:

• Expunging content leaves a trace: Whenever something is expunged, leave a message like “[Thing] has been expunged by [user] on [timestamp].” Bonus points for letting the user doing the expunging add a short comment/reason in addition to the default message. For comments, for example, expunging would simply mark the comment as deleted and replace its content.
• Wherever something can be expunged, users should not be able to expunge their own content. This should be obvious.

I would like to see User Feedback not flooded with bans of sockpuppets and spambots. That persistent spambot Renim got rid of last year left over 100 pages of nothing but bans.